The diagnostic and prognostic role of MMP-1 and MMP-9 plasma concentrations in idiopathic pulmonary fibrosis

Abstract

Background

Idiopathic pulmonary fibrosis (IPF) represents a chronic and irreversible progressive disease, while acute exacerbation of idiopathic pulmonary fibrosis (AE-IPF) represents a catastrophic event in the course of the disease with high mortality. So far, there is no known accurate biomarker with a diagnostic or prognostic role in IPF. Matrix metalloproteases (MMPs), belonging to the family of zinc-containing metallopeptidases, play an active role in extracellular matrix remodeling and in regulating fibrogenesis.

Objectives

The study’s aim was to assess the potential diagnostic and prognostic role of plasma MMP-1 and MMP-9 in IPF and their association to disease progression, acute exacerbation development, and death.

Design

Consecutive stable and exacerbated patients referred to our department from October 2013 until December 2017 were enrolled in the study.

Methods

Demographic, clinical, and laboratory data were collected. MMP-1 and MMP-9 plasma concentrations were measured. All patients were followed-up until death or until 1^st July 2020. We analyzed data using tobit regression analysis, Kaplan-Meier analysis, Cox regression analysis, and fractional polynomials.

Results

One hundred twenty-eight patients were included in the study [74 stable IPF patients, mean age (SD) 71.5 years (8.05) and 54 exacerbated, mean age (SD) 71.8 years (10.15)] as well as 26 healthy controls, mean age (SD) 61.4 years (10.4). At baseline, both stable IPF and exacerbated patients presented higher concentrations of MMP-9 in plasma compared to healthy controls independently of age, sex, and smoking (p=0.006 and p=0.002, respectively). In stable IPF, each 1 µg/L increase in plasma MMP-1 concentration at baseline was associated with an increase in the hazard rate of death by 20% (95%CI: 2% to 42%; adjusted for age, sex, smoking and FVC).

Conclusion

In our patients with stable IPF, MMP-1 concentration is prognostic for death, whereas MMP-9 has diagnostic value for IPF. These results may lead to better patient stratification and personalized therapeutic approaches.

Plain language summary

Idiopathic pulmonary fibrosis (IPF) represents a chronic and irreversible progressive disease, while acute exacerbation of idiopathic pulmonary fibrosis (AE-IPF) represents a catastrophic event in the course of the disease with high mortality. So far, there is no known accurate biomarker with a diagnostic or prognostic role in IPF. The study’s aim was to assess the potential diagnostic and prognostic role of plasma MMP-1 and MMP-9 in IPF and their association to disease progression, acute exacerbation development, and death. In stable IPF, each 1 µg/L increase in plasma MMP-1 concentration at baseline was associated with an increase in the hazard rate of death by 20%. Our results indicate that MMP-1 concentration is prognostic for death, whereas MMP-9 has diagnostic value for IPF.

Introduction

Idiopathic pulmonary fibrosis (IPF) is a chronic, progressive lung disease characterised by scarring of the lung tissue, leading to difficulty in breathing and reduced oxygen exchange, with no known cause.^1–3 It is diagnosed by the presence of a histopathologic and/or radiologic pattern of usual interstitial pneumonia after excluding other known causes of pulmonary fibrosis.^1–4 It remains a disease with an unpredictable clinical course and is often fatal. ⁵ In most patients, disease progression is slow over many years. Some experience a more accelerated course resulting in an end-stage disease early on, whereas others experience episodes of acute respiratory worsening of unknown cause, known as acute exacerbations of idiopathic pulmonary fibrosis (AE-IPF).^5–7 It is estimated that 5% to 10% of IPF patients develop acute exacerbation annually, defined as unexplained worsening of dyspnoea accompanied by severely impaired gas exchange and new radiographic infiltrates, representing a major cause of morbidity and mortality. Approximately 40% of all deaths in IPF are attributed to AE-IPF.^8–11

Current disease pathogenesis in IPF suggests an aberrant wound healing process in response to chronic repetitive injurious stimuli (e.g., environmental toxins, cigarette smoking, viral infections) leading to apoptosis of epithelial cells, activation of neighbouring fibroblasts, expansion to myofibroblasts, which results in fibrogenesis, exaggerated accumulation of extracellular matrix components (ECM) and remodelling.^12–15 Lately, the aberrant proliferation of bronchiolar epithelial cells and the subsequent bronchiolization, that is, the replacement of alveolar epithelium with bronchiolar-like cells, are also considered hallmarks of the disease progression.^16,17 ECM seems to be under constant regulation by several cytokines, growth factors, and proteases. ¹⁸ Disruption of the balance in the regulating mechanisms between ECM formation and degradation leads to excessive accumulation of ECM components. ¹⁹ Among the regulating proteins, metalloproteases (MMPs) are key players in ECM remodelling and pulmonary fibrosis development, acting as positive regulators of fibrogenesis. ²⁰ MMPs are zinc-dependent endopeptidases that act proteolytically on ECM components, regulating crucial repair processes, including cell migration as well as leukocyte and chemokine activation.^20–25 Their classification is based on the discreet substrate specificity, for example, into interstitial collagenases (MMP-1), gelatinases (MMP-9), or matrilysins (including MMP-7).^14,21

Among the genes expressed in IPF lungs, MMP-1 and MMP-9 stand out as being among the most prominently upregulated. ²⁶ MMP-1 is primarily overexpressed in hyperplastic type II alveolar epithelial cells and bronchiolar cells lining honeycomb cysts,^26,27 being responsible for fibrillar collagen degradation by breaking down interstitial collagen types I, II, and III. ²⁸ Moreover, MMP-9 is markedly upregulated within fibroblast and myofibroblast foci, contributing to a fibrogenic feedback loop by either activating latent TGF-β1 or engaging the ERK1/2 signaling cascade.^29,30 This activation of the extracellular signal-regulated kinase (ERK) pathway promotes further fibroblast migration, reinforcing MMP-9’s role as a key pro-fibrotic mediator.^28,31 The involvement of both MMP-1 and MMP-9 in pulmonary fibrosis has been further validated through experimental models of lung fibrosis.^32–34

Idiopathic pulmonary fibrosis, with a gradually increasing incidence and a pathogenesis currently partly elucidated, remains a major clinical challenge. ³⁵ In a previous study, we demonstrated that during AE-IPF, plasma levels of the cytokines IL-6 and IL-8 were significantly elevated compared with those of IPF patients in a stable disease state. ³⁶ Notably, higher circulating IL-6 and IL-8 concentrations were associated with an increased risk of mortality across the entire IPF cohort. ³⁶ In a separate study, we also demonstrated that patients with AE-IPF exhibited significantly shorter telomere length (TL) compared with those with stable IPF. ³⁷ Individuals with TL values below the cohort median (<0.72) had reduced overall survival, independent of age, sex, smoking status, and lung function impairment. ³⁷ These findings further support the growing recognition of telomere length as a key biomarker in IPF, with shorter telomeres consistently associated with poorer clinical outcomes.^37–39 Thus, the identification of reliable peripheral blood biomarkers, whether diagnostic or prognostic, that mirror the underlying pathogenic processes would be a crucial step toward enabling earlier diagnosis and more accurate prediction of disease progression. ³⁵ Unfortunately, so far, none of the investigated biomarkers has been widely used in clinical practice. Based on the above background, we hypothesized that increased MMP-1 and MMP-9 concentrations in plasma would be capable of discriminating patients with IPF from healthy volunteers, and also that their high levels would be associated with poorer survival in IPF. This study aims to investigate the potential role of peripheral blood concentrations of MMP-1 and MMP-9 by measuring their associations at the time of diagnosis of stable IPF and time to death. In addition, we also evaluated potential differences in MMP concentrations between exacerbated and stable patients.

Methods

Study type

The study represents a retrospective cohort study. All patients were invited to participate in the study on the first day of their admission. This study was performed in line with the principles of the Declaration of Helsinki. The reporting of this study conforms to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement (see Suppl File). ⁴⁰

Study population

All consecutive IPF patients referred and followed up in the 2^nd Pulmonary Medicine Department of Attikon University Hospital, Athens, from October 2013 to December 2017 were included. Inclusion criteria included a diagnosis of IPF according to the consensus criteria of the American Thoracic Society/European Respiratory Society/Japanese Respiratory Society/Latin American Thoracic Association, using a multidisciplinary approach for each case. ³ Patients presenting AE-IPF during the study period were considered cases. They fulfilled all the proposed criteria of the international guidelines, including worsening of dyspnoea within the last 30 days, exclusion of identifiable causes, while any new ground glass opacities or consolidations were confirmed by new chest computed tomography (CT) at the time of admission.^3,6,8 We also excluded deteriorating IPF patients secondary to cardiac insufficiency as per echocardiogram and measuring N-terminal prohormone of brain natriuretic peptide concentrations in all exacerbated patients. All patients were followed up regularly until death or until 1^st June 2020, whichever came first. During follow-up examinations, IPF patients were submitted to physical examination, pulmonary function tests, and high-resolution chest computed tomography (HRCT). Also, it was recorded if patients progressed during the disease and if they presented an acute exacerbation episode. Progression was defined as either a decline by more than 10% of Forced Vital Capacity (FVC) or 5% of Diffusing Capacity of the Lung for Carbon Monoxide (DLCO) in one year of follow-up or aggravation of the extent of lung fibrosis in the HRCT of the follow-up one year after enrolment. Finally, 26 healthy individuals visiting hospitalized patients at the same time as the ILD patients in our department were selected as apparently healthy controls.

Clinical data

Information on demographics, medical history, treatment, and smoking status was collected for all participants. Clinical and laboratory findings were recorded, including pulmonary function test results, time since first diagnosis, comorbidities as well and data on any immunosuppressive and/or antifibrotic treatment. In all patients hospitalized for acute exacerbation, computed tomography pulmonary angiogram (CTPA) was performed to exclude pulmonary embolism. Blood, sputum and bronchoalveolar lavage (BAL) cultures, when feasible, were performed to exclude infections. In all patients, the Gender, Age, and Physiology (GAP) Index Score was calculated.

Blood collection and laboratory measurements

In all participants, a blood sample of 5 mL was collected using standardized phlebotomy procedures. In patients with AEs, samples were collected on the first day of admission. All plasma samples were immediately obtained via centrifugation, aliquoted, and frozen at -80 °C. Plasma concentrations of metalloproteases, including MMP-1, MMP-7, and MMP-9, were measured by enzyme-linked immunoabsorbent assay (ELISA) using Milliplex kits according to the manufacturer’s instructions (Millipore, Billerica, MA, USA). MMPs were analyzed on a Luminex 200 platform (Luminex Corp, Austin, TX, USA).

Statistical analysis

Statistical analysis was performed using Stata v17 (StataCorp, College Station, TX, USA).

We constructed a directed acyclic graph (DAG) to identify potential confounders that required adjustment in our statistical analysis (Figure 1). Lung function reference values, including FVC, are traditionally based on anthropometric factors, such as weight, height, sex, and age. ⁴¹ Moreover, FVC is known to decline with age, while age also affects the risk of death, as the probability of dying increases with age. In addition, the circulating MMP concentrations, among them MMP-1 and MMP-9, present age-specific and gender-specific differences. ⁴² Metalloprotease expression also seems to be affected by the amount of cigarette smoking. ⁴³ Furthermore, the association of smoking with the risk of death is well established, as more than 7 million deaths are attributed directly to tobacco use, while around 1.3 million are the result of non-smokers being exposed to second-hand smoke. ⁴⁴ Based on these putative causal relationships, we considered age, sex, and smoking as potential confounders. As shown in Figure 1, FVC should not be included in the adjustment as it is a potential collider, and by blocking this path, a backdoor path is opened through age, gender, and smoking, introducing bias in our analysis.

Figure 1.

Putative causal variables affecting the relationship between peripheral blood MMPs concentrations and death in IPF patients.

We used tobit regression to compare plasma concentrations of MMP-1 and MMP-9 between stable and exacerbated IPF patients. Tobit regression was used to account for MMP concentrations that were left-censored at the detection limits. Again, the differences were adjusted for known potential confounders, including age, gender, and smoking.

We used Kaplan-Meier analyses and Cox proportional hazard models to investigate the association of MMPs (scaled as μg/L) with the time to death in IPF patients. We used as lower cut-points for both MMPs the lower limit of detection (MMP-1: 0.027 μg/L; MMP-9: 0.014 μg/L).

Results

Demographic and clinical characteristics of the study patients

The demographic and clinical characteristics of the study subjects are presented in Table 1. In total, one hundred twenty-eight patients were included in the study [74 stable IPF patients, mean age 71.5 years; and 54 exacerbated, mean age 71.8 years]. Most patients were male (75.8%, or 97/128) and ex-smokers (72.7%, or 93/128). Based on baseline lung function (FVC₁ and DLCO₁) and the GAP index, the extent of pulmonary fibrosis in our stable patients with IPF appeared to be moderate. Specifically, our patients demonstrated physiological impairment within the range usually considered intermediate severity in IPF (FVC 50–70% predicted and/or DLCO 35–50% predicted) or were classified as GAP stage II (4–5 points), as shown in Table 1. In the total cohort, 29% received long-term oxygen therapy (LTOT), while most patients did not receive antifibrotics as the study concerned mainly the era before the widespread use of antifibrotics in clinical practice. Patients were followed up for a mean duration of approximately 2.5 years, while stable patients had a longer follow-up period of a mean duration of 3.5 years, until the 1^st of June of 2020. In Table 2, lung function tests of the participants, including FVC and DLCO at six months of follow-up, as well as the percentage of the patients that progressed during a one-year follow-up, are presented.

Table 1.

Baseline characteristics of the study participants.

​	IPF stable n=74	IPF exacerbation n=54	Controls n=26
Age, years	71.54 (8.1)	71.83 (10.2)	61.4 (10.4)
Sex, male	59 (79.7%)	38 (70.3%)	14 (53.9%)
Smoking
Current smokers	9 (12.2%)	2 (3.7%)	8 (30.8%)
Ex-smokers	49 (66.2%)	44 (81.5%)	8 (30.8%)
No smokers	16 (21.6%)	8 (14.8%)	10 (38.5%)
GAP index	4.18 (1.6)	4.9 (1.45)	n/a
FVC1, %	74.8 (20.6)	63.8 (17.0)	97.9 (11.0)
DLCO1, %	49.3 (19.4)	33.8 (13.8)	n/a
Treatment
None	71 (95.6%)	45 (83.3%)	26 (100%)
Antifibrotics	1 (1.35%)	0 (0%)	0 (0%)
Corticosteroids	2 (2.7%)	9 (16.7%)	0 (0%)
LTOT	18 (24.3)	37 (68.5)	0 (0)
Lung surgical biopsy	2 (2.7%)	2 (3.7%)	n/a

Values indicate mean (SD) or n (%). FVC₁: Forced Vital Capacity; DLCO₁: Diffusing Capacity of the lung for carbon monoxide at baseline; FVC₂: Forced Vital Capacity after 6 months follow-up; DLCO₂: Diffusing Capacity of the lung for carbon monoxide after 6 months follow-up; GAP: The Gender, Age, and Physiology Index Score [Maximum possible points=8]; LTOT: long-term oxygen therapy, n/a: not applicable.

Table 2.

Follow-up characteristics of the study participants.

​	IPF stable n=74	IPF exacerbation n=54	Controls n=26
FVC2, %	74.0 (19.7)	63.5 (15.2)	82.6 (15.8)
DLCO2, %	48.5 (19.0)	33.6 (15.0)	n/a
Progression, n (%)	18 (24.3%)	n/a	n/a

Values indicate mean (SD) or n (%). FVC₂: Forced Vital Capacity after 6 months follow-up; DLCO₂: Diffusing Capacity of the lung for carbon monoxide after 6 months follow-up; Progression: as explained in Methods over 1 year of follow-up; n/a: not applicable.

Plasma concentrations of metalloproteases

We used histograms to visually inspect the distribution of metalloprotease (MMP) concentrations (Figure 2).

Figure 2.

Distribution of observations regarding MMP-1, and MMP-9.

The lower limit of detection was at 0.027 μg/L for MMP-1 and 0.014 μg/L for MMP-9, and variables were left-censored to the above lower limits.

For MMP-1 and MMP-9, 10.4% and 1.3% of the total values were below the limit of detection, respectively. The summary of non-censored values was n=154, median (25-75 percentiles): 1.62 μg/L (0.76-3.18 μg/L) for MMP-1; n=154, median (25-75 percentiles): 7.03 μg/L (51.81- 102.68 μg/L) for MMP-9. The summary of censored values is presented below (n=138, median (25-75 percentiles) 1.8 μg/L 6 (0.95- 3.35 μg/L) for MMP-1 and n=152, median (25-75 percentiles) 70.87 μg/L (52.70-102.78 μg/L) for MMP-9).

In Table 3, the mean (SD) plasma concentrations of MMP-1 and MMP-9 are presented by patient group. There was no evidence that both MMP-1 and MMP-9 plasma concentrations differed between IPF stable and IPF exacerbation patients (Figure 3). However, both IPF stable and exacerbation patients had higher MMP-9 plasma concentrations compared to healthy controls (Figure 3).

Table 3.

Plasma concentrations of MMP-1 and MMP-9, by patient group.

​	IPF stable n=74	IPF exacerbation n=54	Controls n=26
MMP-1 (μg/L)	1.30 (2.054)	2.99(4.18)	1.24 (1.24)
MMP-9 (μg/L)	79.04 (40.65)	84.00 (52.50)	58.40 (24.29)

Values indicate mean (SD).

Figure 3.

Baseline plasma MMP1- and MMP-9 concentrations among IPF stable, exacerbated patients, and controls.

a. MMP-1; The Y-axis range is 0 to 20, and one outlier for IPF exacerbation is not shown (27.83 μg/L).

b. MMP-9; The Y-axis range is 0-250.

Kaplan-Meier survival curves

Plasma MMP-1 and MMP-9 and the odds of death

IPF stable patients with elevated concentrations of plasma MMP-1, had reduced time to death compared to their peers with low concentrations (p=0.05), as shown in Figure 4. Regarding MMP-9, there was also an indication for poorer survival among patients with higher concentrations (p=0.21, Figure 4).

Figure 4.

Kaplan-Meier curves for IPF stable patients, by plasma MMPs concentration category

*The cut-point to dichotomize the groups was the mean of the MMPs concentrations, which was 2.26 μg/L for MMP-1 and 77.41 μg/L for MMP-9, respectively.

In addition, regarding all IPF patients, there might be also an indication for poorer survival among patients with higher concentrations of MMP-1, (p=0.126, Figure 5).

Figure 5.

Kaplan-Meier curves for all IPF patients, by plasma concentration category

*The cut-point to dichotomize the groups was the mean of the MMPs concentrations, which was 2.26 μg/L for MMP-1 and 77.41 μg/L for MMP-9, respectively.

Finally, there was no evidence that patients with AE-IPF at baseline with higher concentrations of MMP-1 and MMP-9 in plasma had reduced survival compared to their peers with low concentrations (p=0.611 and p=0.151, respectively), as shown in Suppl. Figure S1. In Cox regression analysis, when adjusting for age, sex, smoking and FVC, there was no evidence that the proportional hazard assumption was violated (p=0.19). In stable IPF patients, an increase in the concentrations of MMP-1 by 1μg/L increased the hazard rate of death by 20%, whilst there was no evidence of association between plasma MMP-9 and the hazard of death (Table 4; estimates adjusted for age, sex, smoking and FVC). No support for association was detected between plasma concentrations of MMP-1 and MMP-9 and the risk for progression and AE-IPF.

Table 4.

Associations of plasma concentrations of MMPs with death rates in patients with stable IPF at baseline (n=74).

IPF stable at baseline	Hazard ratio (95% CI)	p-value
Crude analysis
MMP-1 (μg/L)	1.32 (1.13-1.55)	0.001
MMP-9 (μg/L)	1.01 (1.0-1.01)	0.200
After adjusting for age, sex, and smoking
MMP-1 (μg/L)	1.23 (1.03-1.46)	0.020
MMP-9 (μg/L)	1.00 (0.99-1.01)	0.620
After adjusting for age, sex, smoking and FVC
MMP-1 (μg/L)	1.20 (1.02-1.42)	0.030
MMP-9 (μg/L)	1.00 (0.99-1.01)	0.532

In addition to confirming our previous results showing a potential association of MMPs plasma concentrations with the risk of death in IPF, we designed a model with fractional polynomials that provided further evidence of such a link (Tables 5 and 6).

Table 5.

Associations of plasma concentrations of MMPs with hazard rates of death, adjusting for age, sex, and smoking in IPF stable patients at baseline^*.

IPF stable at baseline	MMP-1 (μg/L)	MMP-9 (μg/L)
Hazard ratio	1.21	1.00
95% C.I.	1.31- 1.43	0.99-1.01
p-value	0.020	0.640

*For age, we used fractional polynomials to allow a non-linear relationship with the log-hazard ratio.

Table 6.

Associations of plasma concentrations of MMPs with hazard rates of death, adjusting for age, sex, smoking and FVC in IPF stable patients at baseline^*.

IPF stable at baseline	MMP-1 (μg/L)	MMP-9 (μg/L)
Hazard ratio	1.20	1.00
95% C.I.	1.02- 1.42	0.99-1.01
p-value	0.030	0.532

*For age, we used fractional polynomials to allow a non-linear relationship with the log-hazard ratio.

Moreover, there was no evidence of association between plasma MMP-1 and MMP-9 concentrations and the hazard of death for AE-IPF patients at baseline, as well as when we included all IPF patients at baseline (Suppl. Tables 1 and 2).

Discussion

At presentation, both stable and exacerbated IPF patients exhibit higher concentrations of MMP-9 compared to healthy controls, providing additional evidence regarding its value in IPF. In addition, we found plasma MMP-1 concentrations to be associated with an increased hazard rate of death in IPF, independently of age, sex, smoking and FVC. Indeed, IPF patients with higher plasma MMP-1 concentrations present poorer survival compared to those with lower concentrations.

Our findings on the diagnostic role of MMP-9 in distinguishing patients with IPF, whether stable or experiencing an exacerbation, from healthy controls further support the importance of MMP-9 in IPF. Previous basic and clinical studies have underlined the role of MMPs in the pathobiology of IPF.^27,45 In experimental lung fibrosis in mice, MMP-9 protein levels are found to be increased in bronchoalveolar lavage fluid and lung homogenates of mice with adenovector-mediated gene transfer of active transforming growth factor-beta1 (AdTGF-β1)-induced fibrosis. ⁴⁵ MMP-9 promotes lung fibrosis by releasing the complex, formed by the latency-associated protein (LAP) that is covalently bound to the latent TGF-β binding protein (LTBP), increasing in this way the active form of TGF-β1.^46,47 In addition, other studies using the bleomycin model have also shown higher MMP-9 protein concentrations in rats and mice.^48,49 Regarding the clinical data, our results are in line with previous reports that have proposed a role of MMP-9 in IPF.^20,21 More precisely, MMP-9 increased concentrations have already been detected in the peripheral blood of IPF patients,^27,50–52 but also in the lung tissue of pulmonary fibrosis patients with usual interstitial pattern, either attributed to IPF or to non-IPF.^51,53,54

Therefore, our findings of the observed increased MMP-9 plasma concentrations may reflect the underlying chronic and dysregulated ECM remodelling, seen in IPF. ⁵¹ MMP-9 is known to participate in the ECM degradation, contribute to fibroblast activation and further promote the fibrotic feedback loop. Indeed, we revealed that both stable and exacerbated IPF patients present higher concentrations of MMP-9 in plasma compared to healthy controls, after adjusting for age, sex, and smoking. Our results are in line with previous works that have shown that, among other MMPs, MMP-9 and not MMP-1 is a top candidate for distinguishing patients with IPF from controls in multivariable analyses. ⁵¹ Particularly, the IPF-PRO Registry found that MMP-9 levels were significantly higher in IPF patients than in healthy controls. ⁵¹ In contrast, in a previous study, Rosas IO and colleagues suggested that MMP-1, along with MMP-7, could be key elements of a peripheral blood protein signature for IPF, but it has yet to be routinely applied in clinical practice for IPF monitoring. ²⁷ Unlike these findings, our results highlight the diagnostic potential of MMP-9 specifically for detecting IPF patients, while we could not find a difference in MMP-1 concentrations between stable patients and controls. Indeed, Rosas IO et al. showed that the diagnostic potential of MMP-1 became significant only when elevated alongside MMP-7, suggesting a synergistic or additive effect. ²⁷ In other words, although MMP-1 levels were elevated in IPF, it was MMP-7 that played the more consistent and robust role as a standalone biomarker. In addition, we found no evident difference in concentrations of either MMP between stable and exacerbated patients. This interplay may hint at the complex pathobiology in the context of AE-IPF where multiple matrix metalloproteinases and mediators may contribute to disease phenotype. One could speculate that MMPs concentrations may reflect the ongoing, chronic underlying pathology in IPF (epithelial injury, aberrant repair and ECM remodelling), and not the acute episodic biology of AE-IPF, characterised by diffuse alveolar damage and neutrophilic inflammation, dominated probably by cytokines, and coagulation pathways rather than ECM-remodeling enzymes.^{6,11,14,27,36} In general, MMPs are considered baseline disease severity biomarkers and not acute-phase discriminators.^14,27

Regarding the prognostic value of MMPs, our study reveals a strong link between elevated plasma MMP-1 levels and higher mortality risk in IPF, regardless of age, gender, smoking status and FVC. Notably, IPF patients with increased plasma MMP-1 concentrations exhibit worse survival outcomes compared to those with lower levels. The prognostic role of MMP-1 in IPF is less clearly defined compared to other matrix metalloproteinases like MMP-7, but some insights have emerged from recent studies.^20,51,55 So far, MMP-7, the most extensively studied metalloprotease, has been shown to predict poor overall survival.^56,57 Therefore, we provide evidence that MMP-1 may also represent a reliable and useful prognostic biomarker, capable of stratifying IPF patients with increased risk of death. As MMP-1 represents a collagenase involved in the extracellular matrix turnover, and its expression in IPF lungs is primarily localized to epithelial cells and not fibroblastic foci, plasma MMP-1 concentrations may predict outcomes, reflecting epithelial injury and aberrant extracellular matrix remodelling, both key biological processes driving disease progression. ⁵⁵ Moreover, MMP-1 is also part of a validated peripheral blood protein signature associated with more severe disease biology, and its elevation correlates with physiological impairment and disease severity, both of which may translate into higher mortality risk. ⁵⁵ As higher MMP-1 levels correlate with worse disease biology, reflecting the ongoing dysregulated ECM remodelling, elevated peripheral blood MMP-1 may also be associated with an increased risk of mortality. Overall, identifying novel diagnostic and prognostic peripheral blood biomarkers, indicators of disease activity, such as MMP-1 and MMP-9, remains indeed so crucial in the management of IPF, as that would facilitate prompt diagnosis and optimize disease monitoring.

Our study has several limitations. Although we observed strong evidence of an association, the relatively small sample size limits the generalizability of our findings and may affect the precision of the estimated effect. Another might be the lack of simultaneous measurements of MMP-1 and MMP-9 in bronchoalveolar lavage and tissue in parallel with those in plasma. However, we focused on peripheral blood detection, as we believe that biomarkers are preferable to be easily sampled, like in blood, to achieve widespread use in clinical practice. Another issue in our study was that several values of MMP-1 and MMP-9 were below the detection limits, therefore, we had to handle our data as censored and use tobit regression analysis, which is the appropriate statistical model in such cases. Moreover, as blood collection and patients’ enrolment were conducted before the widespread introduction of antifibrotics in clinical practice, our patients are naïve without receiving either antifibrotics or immunosuppressives, a fact that might have permitted us to have clearer results regarding the potential role of MMP-1 and MMP-9 during the clinical course of IPF. In addition, the current use of antifibrotic therapy as standard of care in IPF does not appear to affect the applicability of our findings regarding the diagnostic and prognostic role of MMP-1 and MMP-9. To date, there is no evidence of a consistent direct effect of antifibrotic therapy (pirfenidone and nintedanib) on circulating MMP-1 or MMP-9 levels in IPF. However, there might be an effect of nintedanib and not pirfenidone on downstreaming ECM-related neoepitopes, but not MMPs themselves, an effect reflecting changes in ECM turnover, rather than any direct MMP inhibition.^57,58 Our results need further validation because our study represents a single-academic-centre study based on a well-defined group of IPF patients.

Conclusions

In our patients with stable IPF, MMP-1 concentration is prognostic for death, whereas MMP-9 has diagnostic value for IPF. These results may lead to better patient stratification and personalized therapeutic approaches.

ORCID iD

Ioannis Tomos https://orcid.org/0000-0002-9978-4823

Ethical considerationss

The study protocol was approved by the local Ethics Committee of Attikon University Hospital, and participants provided written informed consent (201/23-4-14).

Consent to participate

All patients were invited to participate in the study on the first day of their admission. This study was performed in line with the principles of the Declaration of Helsinki.

Consent for publication

Patients have approved their participation in the study and the publication of its results.

Data Availability Statement

The datasets generated during and/or analyzed during the current study are available from the corresponding author on request.*